I have trade volume daily data for 12 years of 500 companies. I wanna divide these companies into high, medium and low volume portfolio using trade volume data. Can someone help me in this regard
Hi Komal,
I have in mind two easy ways. First (the most simple), you can use distribution quartiles. For example, companies with low volume are those below the first 25% of the distribution, while those with high volume are above 75% (the others have medium volume). Alternatively, you can use time series cluster analysis approaches to define 3 groups of companies with homogenous volume's time patterns. In this way, you can let the data decide the composition of the 3 groups without making any questionable choices.
Yes but I don't understand one thing that is I need only one value for the one company for while time period but I have so many observation so how can I get that one value for one company which represent entire time period... because to divide them in quartile I need one series for trade volume with 500 observation...
You can simply use the sum of the volumes over the entire time period or the average value! However, I would always prefer doing cluster analysis.
@ Raffaele Mattera will you please tell me what type of cluster analysis I use in this because there are many types of time series cluster analysis like hierarchical or k- means....
- Badges
- Science topic
- Similar topics
- Mathematics
- Algebra
- Matrix